1. Field of the Invention
The invention relates to a readout circuit, more particularly to a readout circuit adapted for receiving a line analog image signal from a sensor array of an image sensor.
2. Description of the Related Art
Complementary Metal Oxide Semiconductor (CMOS) image sensors are widely used in digital image capturing devices, such as video camcorders, digital cameras, camera cell phones, web cameras, etc. As shown in FIG. 1, a conventional image sensor 1 generally includes a sensor array 10 including a matrix of sensor units (not shown). Each of the sensor units is capable of sensing light, and is operable to output a sequence of line analog image signals (a1) that correspond to the light sensed thereby. The line analog image signals (a1) are outputted to an amplifying unit 111 of a readout circuit 11 for amplifying and correcting amplitude of the line analog image signals (a1). An n-bit analog-to-digital converter 112 of the readout circuit 11 is coupled to the amplifying unit 111, and converts the amplified and corrected analog image signal (a2) from the amplifying unit 111 into a corresponding n-bit digital image signal (d1), which is subsequently outputted to a back-end circuit (e.g., a digital signal processing (DSP) unit) for signal processing.
During analog-to-digital conversion in the readout circuit 11 of the prior art, signal distortion is likely to occur. As shown in FIG. 2, ideal image signal voltage distribution (E1) ranges from 0 to V1. However, actual image signal voltage distribution (E2) differs from the ideal image signal distribution (E1) by an offset value, Voffset. In other words, the actual image signal voltage distribution (E2) ranges from Voffset to V2, where V2=V1+Voffset. It has been proposed heretofore to use an analog adder-subtracter (not shown) to subtract a modifying signal Vp from the actual image signal in an attempt to reduce image distortion. However, since the adder-subtracter has a higher resolution than that of the analog-to-digital converter, the modifying signal Vp differs from the offset value Voffset, i.e., Vp≠Voffset. Therefore, although image distortion is reduced in the distribution (E3) of the modified image signal, it is not completely eliminated, resulting in a relatively low capturing resolution of the image.